About Harmonova 01
WE REBUILT CUSTOMS DATA FROM THE SOURCE.
Harmonova is a real, open, MIT-licensed customs database, with every tariff line traced back to the official government schedule it came from. Not an AI wrapper: a living system that re-checks every source on a schedule, and shows its work.
Harmonova HV-1 · power it on, watch it boot
Why we built it 02
Customs data deserved better.
Built by a founder with 10+ years in ecommerce and logistics, tired of how complex customs had become and how poor the data was. So we built it.
“Customs should be the easy part. The data just had to be right, and sourced, so people could trust it.”
How the database is built
Built from the official feeds, kept current.
Five principles behind the database. Global coverage at the 6-digit level, full national depth where it matters, and a source on every line.
- 01
Every 6-digit heading on Earth
The full WCO Harmonized System, every country covered at the 6-digit level.
5,635WCO headings - 02
Full 10-digit depth where it matters
Complete national schedules for the US, EU and UK, pulled from official sources.
19,147US23,000+EU16,000+UK - 03
A source link on every line
Each code carries the official schedule entry it was pulled from, never an averaged rate.
- 04
Re-synced daily
Reconciled against the official feeds every day, so terminations and new measures flow straight through.
- 05
A receipt on every lookup
Every result returns the code, rate, source and a confidence score, reproducible and audit-defensible.
Coverage 03
Every heading on Earth, sourced.
The full WCO Harmonized System at the international 6-digit level, every trading country covered, with complete national depth in the markets that clear the most trade. Pulled straight from the official government feeds.
- USITC HTSUnited States
- TARICEuropean Union
- Trade TariffUnited Kingdom
- WCO HSEvery country
Under the hood 04
Built to be audited, not taken on faith.
If you are bringing Harmonova in past a security review or a technical evaluation, here is exactly what it is, what it runs on, and why a language model on its own cannot replace it.
THE STACK
Open source, top to bottom
Purpose-built from a blank page, not bolted onto a decades-old tax suite. A TypeScript API over Postgres, deployed on Fly, where one clean schema unifies the WCO base and the national schedules and every line carries its source. The database, the classifier, and the source links live in one MIT-licensed repository you can read, self-host, or fork. No black box to take on faith.
THE DATA PIPELINE
Parsed from the source, synced daily
The database is sourced, not generated. Governments publish their tariff schedules as official data, files and live feeds built for software, not web pages built for people. We pull from those directly (the WCO Harmonized System, the USITC HTS data export, EU TARIC, the UK Trade Tariff feed), parse them line by line into one normalised schema, and re-pull every day. That is the opposite of scraping: no guessing data out of a rendered page, and a model never writes a tariff code, it only maps your product to one.
THE CLASSIFIER
Grounded, not guessing
Classification runs on Anthropic's Claude at temperature 0, wrapped in a deterministic category guard and a curated alias layer that catch the cross-chapter mistakes a raw model makes. Every output is schema-validated, and it never invents a code that is not in the database.
SECURITY & PRIVACY
Built to pass a review
Postgres over enforced TLS, scoped API keys, rate limiting, and standard hardening. Catalog data sent for classification is not used to train models, and you can self-host the entire stack if your policy requires it.
The question every technical buyer asks
“Can't anyone just build this with AI?”
Honestly? With months of work, maybe. But you would not be building an AI. You would be building a database.
Harmonova is not a model that guesses. It is a real, open, MIT-licensed customs database, built from the ground up, with every one of 63,000+ tariff lines linked back to the official government schedule it came from. That is months of mapping, auditing and daily syncing, not a weekend of prompting.
Ask a chatbot for an HS code and it hands you a confident answer that fails an audit, because it never read the tariff schedule. Plenty is possible with AI; a defensible classification is not, without the verified, sourced data and the customs judgment underneath it.
The moat was never the model. It is the work, the sourced data, and the expertise behind every answer. We already did it, so your team does not have to.
How it is different from the incumbents
The established customs and tax suites (the Zonos and Avalara tier) are closed, priced for the enterprise, and rarely show you where a given rate came from. Harmonova is open-source and MIT-licensed, API-native and built for a developer to drop in, and every single line traces back to its official government source. You are not trusting a vendor's black box. You are trusting the published schedule, with a link to prove it.